Enzyme-linked immunosorbent assay value for BP180 (BP antigen II) (A) and BP230 (BP antigen I) (B) autoantibodies vs patient age. Smooth curves were fit separately for men (solid line) and women (dotted line). Dashed line shows the cutoff value of 9 U/mL for a positive result (≥9 U/mL).
Wieland CN, Comfere NI, Gibson LE, Weaver AL, Krause PK, Murray JA. Anti–Bullous Pemphigoid 180 and 230 Antibodies in a Sample of Unaffected Subjects. Arch Dermatol. 2010;146(1):21-25. doi:10.1001/archdermatol.2009.331
Copyright 2010 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2010
To evaluate the prevalence of autoantibodies against 2 hemidesmosomal proteins typically found in patients with bullous pemphigoid (BP), BP antigen II (BP180) and BP antigen I (BP230), in persons without BP.
Academic medical center.
An age- and sex-stratified, random, population-based sample of local county patients seen during 2007: 20 men and 20 women per decade of age (from age 20 to 89 years) and 57 patients (33 women and 24 men) aged 90 to 99 years.
Stored serum samples were retrieved for analysis by enzyme-linked immunosorbent assay and indirect immunofluorescence.
Main Outcome Measure
Presence of circulating autoantibodies to BP180 and BP230.
Of the 337 study patients, 25 (7.4%) were positive for 1 or both autoantibodies; these 25 samples all tested negative with indirect immunofluorescence. Autoantibody levels did not vary by age or sex.
Bullous pemphigoid has a higher incidence in the elderly population, but the prevalence of antibodies to BP180 and BP230 did not increase significantly with age or vary by sex in this population-based sample. Other exogenous factors may affect the development of these autoantibodies in a population without clinically evident immunobullous disease, including limitations inherent to the test (false-positive rate).
Bullous pemphigoid (BP) is an acquired autoimmune bullous disorder characterized clinically by tense subepidermal bullae arising on normal or erythematous skin. The pathogenesis of BP has been characterized by circulating autoantibodies directed against basement membrane zone hemidesmosomal proteins.1 Diagnosis is established by a combination of studies including histopathologic assessment by routine microscopy, direct immunofluorescence studies to demonstrate linear deposition of in vivo bound antibodies (IgG and/or C3), and indirect immunofluorescence (IIF) to detect circulating autoantibodies against basement membrane zone proteins. Most commonly, IgG (rarely IgA, IgM, and IgE) antibodies bind to the hemidesmosomal proteins BP antigen I (BP230), a 230-kDa intracellular component of the hemidesmosomal plaque, and BP antigen II (BP180), a 180-kDa transmembrane protein.2- 5 Antibodies against BP180 most commonly react with epitopes in the 16th noncollagenous domain (NC16a).6,7 Serum levels of autoantibodies to BP180-NC16a have been shown to parallel disease activity.8 Furthermore, subepidermal blisters have been reproduced in animal models and in cryosections of human skin when they were exposed to serum samples of patients with BP, which confirms the pathogenicity of BP180 autoantibodies.9,10
Epidemiologic studies show that BP has a higher incidence in the elderly population. A study of the incidence of BP in 3 French populations found the mean age at the onset of BP to be 82 years, and a Polish study found the mean age to be 69 years for women and 67 years for men.11,12 These findings parallel those of a study of the German population, which showed that the risk of BP rapidly increases in patients older than 60 years, with patients older than 90 years having the highest incidence of the disease.13
Abundant evidence supports the presence of autoantibodies to BP180 and BP230 in patients with BP, along with the established pathogenicity of the autoantibodies and their relationship to disease activity. Little is known, however, about the prevalence of BP autoantibodies in the general population. Therefore, our main objective was to determine the prevalence of these autoantibodies in a population-based sample of patients without underlying immunobullous disease. Studies have shown an enzyme-linked immunosorbent assay (ELISA) for BP180 and BP230 autoantibodies to be both sensitive and specific for the disease. We used this ELISA technique to examine serum samples of patients without BP for the presence of autoantibodies against BP180 and BP230.
This study was approved by the Mayo Clinic Institutional Review Board. The source population for the study was patients residing in Olmsted County, Minnesota, who were seen at Mayo Clinic in Rochester, Minnesota, during 2007 and had a serum sample taken.
Study participants were identified from a “waste-blood” list that is compiled daily, which includes all excess serum samples originally collected for clinical purposes but subsequently made available for research use. We used a series of previously developed automated programs to search the waste-blood lists for subjects meeting the study criteria.
The study included Olmsted County residents who had provided research authorization and were aged 20 to 99 years at the time of sample collection. Persons younger than 20 years and those with a history of clinically diagnosed celiac disease or who previously had a blood sample taken as part of clinical evaluation for celiac disease were excluded from the study. The HICDA (Hospital International Classification of Diseases Adapted) codes for all samples were subsequently reviewed to exclude all cases of diagnosed pemphigoid or pemphigus syndromes.
For all identified patients meeting the study criteria, stratified sampling was applied to ensure age and sex balance. The final study sample was assembled by randomly selecting 40 patients (20 women and 20 men) per decade of age from 20 to 89 years and 57 patients (33 women and 24 men) aged 90 to 99 years. For these 337 patients, serum samples were retrieved for analysis. These serum samples had been stored, were no longer needed after being held in the central laboratory for the standard 6 days, and were still available for testing (waste serum samples).
Commercially available BP180-NC16a and BP230 ELISA kits (Medical and Biological Laboratories Co Ltd, Nagoya, Japan) were used to test the serum samples for the presence of autoantibodies to the BP antigens. Reactivity to BP180 and BP230 was analyzed by ELISA using purified recombinant forms of BP180-NC16a and BP230 (amino and carboxy terminals) antigens, as described previously.7 A cutoff ELISA value of 9 U/mL was used (≥9.0 U/mL being a positive result) on the basis of a prior study that assessed the usefulness of commercially available BP180-NC16a ELISA for the initial serodiagnosis of BP.14 That study reported a sensitivity of 89% and a specificity of 98% for this cutoff value. Multiple studies have demonstrated similar sensitivities and specificities for the serodiagnosis of BP at the cutoff value of 9 U/mL using commercially available ELISA kits for detecting BP180 and BP230 autoantibodies.
Serum samples found to be positive for BP180, BP230, or both autoantibodies were further tested with IIF. Indirect immunofluorescence was performed with monkey esophagus as a tissue substrate, which was overlaid with the serum samples, washed, and subsequently overlaid with fluorescein isothiocyanate–tagged antihuman IgG. Antibody titer is determined using monkey esophagus with serial dilutions of the serum samples. This method will detect the presence of circulating antibodies against basement membrane zone components including BP180 and BP230 (Michael J. Camilleri, MD, unpublished data, 2009).
The relationship of autoantibody levels to age and sex was assessed graphically with scatter plots. A smooth curve was fit to each scatter plot to summarize the pattern separately for men and women. The smoothed curves were generated by fitting a nonparametric smoother to the data using a cubic spline routine. Given that the distributions of the autoantibody levels were positively skewed, logarithmic transformations were applied and the average results were summarized using the geometric mean. General linear regression models were fitted to assess the relationship of autoantibody levels (after logarithmic transformation) to age and sex. All analyses were performed using the SAS software package, version 9.0 (SAS Institute Inc, Cary, North Carolina). P < .05 was considered statistically significant.
Of the 337 patients without known bullous disease who were included in the sample, 25 (7.4%; 95% confidence interval, 4.6%-10.2%) had positive results by ELISA (≥9 U/mL) for autoantibodies against BP180 only (n = 11), BP230 only (n = 11), or both (n = 3) (Table). Neither sex nor age was significantly associated with a positive test result (χ2 test, P = .24; 2-sample t test, P = .88). Among these 25 patients with positive results, 15 (60.0%) were men and the mean (SD) age was 61.2 (24.9) years. Among the 312 patients with negative results, 149 (47.8%) were men and the mean (SD) age was 62.0 (23.1) years. The incidence of a positive result was 7.4% among all patients, 7.3% among patients younger than 60 years, and 7.6% among patients 60 years or older.
BP180 and BP230 autoantibody levels did not appear to vary significantly with age or sex (Figure). Levels of both antibodies remained stable across the age groups. After adjusting for age, the geometric mean of the levels did not vary significantly by sex (women vs men, BP180, 1.44 vs 1.56 U/mL [P = .47]; BP230, 1.47 vs 1.52 U/mL [P = .77]).
To further rule out a history of another autoimmune bullous disorder or undetected BP, all 25 ELISA-positive serum samples were tested with IIF, and the medical records of those patients were reviewed for any potentially relevant data that could suggest an underlying clinically occult autoimmune bullous disorder. All serum samples showed negative IIF results (Table). Antibodies to the NC16a domain of BP180 are seen in herpes gestationis as well as BP.15,16 Therefore, the clinical histories of female patients with positive BP180 or BP230 ELISA results were reviewed for a history of pregnancy or use of hormone therapy. In this subgroup, nothing in the history suggested the presence of clinical or laboratory features typical of herpes gestationis. One woman had a newly diagnosed connective tissue disorder. Several patients had a reported history of dermatitis, but none of these cases were noted to be urticarial or suggestive of BP. In addition a review of the patient-completed family history form available for 24 of 25 patients did not reveal any family history of immunobullous disease, including BP or herpes gestationis. Information in the medical history that was considered irrelevant was reviewed but excluded from the Table.
To our knowledge, this is the first study to examine the prevalence of anti-BP180 and anti-BP230 antibodies in a nondisease (non-BP) population-based sample of patients stratified by age and sex. Although studies have shown a higher incidence of BP in the elderly population,11 our study showed that the prevalence of BP180 and BP230 autoantibodies does not appear to vary by sex or to increase significantly with age in a sample of the general population. The test characteristics remained flat across the age groups.
Indirect immunofluorescence has been considered by many to be the diagnostic standard for BP, with reported sensitivities between 75% and 96%.2,17 However, confirmation of disease requires correlation with the clinical pattern, routine pathologic analysis, and IIF to ensure the highest level of accuracy in diagnosis. None of our 25 patients with positive ELISA results have clinical evidence of BP development to date, although it is possible that the disease may develop in the future.
Results of the BP180-NC16a ELISA have been positive in greater than 90% of patients with BP and negative in disease control and normal control serum samples.7,15 Further studies have shown ELISA to be a useful diagnostic tool for the diagnosis of BP. In 23 newly diagnosed cases, Chan et al2 showed ELISA for the anti–BP180-NC16a antibody to have a sensitivity of 96% and to be comparable to traditional diagnostic techniques such as direct immunofluorescence and IIF. A larger study of 102 patients with BP found similar results, with a sensitivity of 89% and a specificity of 98%.14 A study by Thoma-Uszynski et al18 also gave similar results; BP180 ELISA testing showed an overall sensitivity of 95% and specificity of 94%. In that study, nearly 80% of patients had concordant BP180 and BP230 reactivity, and ELISA for BP230 alone had a sensitivity and a specificity of 82% and 65%, respectively. In addition, unpublished data (Carmen J. Rinaldi, MD, and Michael J. Camilleri, MD; 2006) from our institution suggest that ELISAs for BP180 and BP230 autoantibodies are sensitive and specific tests that should be included in the initial screening for diagnosis of BP. These reported sensitivities and specificities correlate with our incidence of 7.4% (95% confidence interval, 4.6%-10.2%) as the false-positive rate.
One might expect that if BP was caused by the presence of autoantibodies alone, a higher incidence of these antibodies would be found in the older age groups. Overall, the incidence of positive ELISA results was 7.3% among patients younger than 60 years and 7.6% among patients 60 years and older. From this, we can speculate that secondary triggering or initiating factors may be involved in the induction of BP in susceptible persons. In a multicenter case-control study by Bastuji-Garin et al,19 the long-term use of aldosterone-antagonist diuretics (a strong association) and the use of neuroleptics (a weak association) emerged as potential risk factors for the development of BP. A higher incidence of neurologic disorders associated with severe functional impairment and poor prognosis, including dementia, cerebral stroke, and Parkinson disease, was noted in a series of 341 consecutive patients with BP.20 A literature review identified 27 patients with radiation therapy–induced BP; most of these patients had development of BP after a median dose of 50 Gy (to convert to rads, multiply by 100), and it predominantly occurred in those with breast cancer.21
In studies seeking other associations with the development of BP (eg, drugs, infection, neurologic disorders), certain confounding factors must be taken into account with respect to the age of the population in which BP most commonly occurs. This group is more likely to be taking multiple medications, including diuretics, and is more predisposed to dementia by virtue of their older age. Hence, caution should be exercised in the interpretation of any identified associations.
A limitation of our study is the possible existence of additional antibody targets besides BP180 and BP230 that may be necessary for the development of BP. Much focus has been placed on the BP180 autoantibody reactivity to the NC16a domain. However, Mariotti et al22 identified a subgroup of BP180-NC16a–negative patients in whom antibodies reacted with different BP180 epitopes. Perhaps a greater number of elderly patients have antibodies to other epitopes or additional antibody targets completely different from BP180 and BP230. In addition to a lack of difference in antibody profiles among decades, sex differences were not seen in our study. Because antibodies to the NC16a domain of BP180 are seen in herpes gestationis (pemphigoid of pregnancy) as well as BP,15,16 an increased prevalence of antibodies may be expected among women. Because we did not find this result, a role for potentially undiagnosed cases of herpes gestationis or a hormonal association in our patients with positive ELISA test results is not likely. Pemphigoid of pregnancy is a rare disease, and its confirmation requires careful clinicopathologic correlation.
The incidence rate of 7.4% might represent the false-positive rate of the test in our hands, but it also could constitute an aggregate of false-positive and true-positive test results, which did not change significantly across the age groups. In contrast, the prevalence of BP is so low that the prevalence of autoantibodies may increase significantly, but the number of cases studied was too small or the study was not sufficiently powered to measure this change. The latter explanation is more likely and represents a significant limitation of this study, given the high sensitivity of the ELISA test, low incidence of BP, and the total number of cases. Despite the high levels of sensitivity and specificity reported for the BP180 and BP230 ELISAs, the ultimate and most accurate diagnosis of BP depends on careful correlation of the clinical findings with histopathologic, immunopathologic, and serologic features. Patients who have negative results in any of these parameters will present challenges to diagnosis and may require close clinical follow-up with serial testing to establish the diagnosis. As we learn more about BP susceptibility and induction patterns, additional factors, other than the presence of anti–basement membrane antibodies, that have a key role in disease pathogenesis will emerge. These factors could then potentially be used to aid in diagnosis, particularly in early stages of disease.
Ultimately, our findings raise questions for future scientific study. A multifactorial pathogenetic model for BP could be considered, one that requires not just the presence of specific pathogenic autoantibodies but also additional triggers (eg, environmental factors, comorbid conditions, drug exposures, concurrent or preceding infections) that may have a role in disease initiation within a susceptible person. In addition, there may be as yet unidentified antigenic epitopes that are necessary for disease development and that the current ELISA tests are unable to identify.
Correspondence: Nneka I. Comfere, MD, Division of Laboratory Dermatology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (firstname.lastname@example.org).
Accepted for Publication: May 2, 2009.
Author Contributions: Dr Comfere had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Comfere, Gibson, and Krause. Acquisition of data: Wieland, Comfere, and Murray. Analysis and interpretation of data: Wieland, Comfere, Gibson, and Weaver. Drafting of the manuscript: Wieland, Comfere, Gibson, and Weaver. Critical revision of the manuscript for important intellectual content: Comfere, Gibson, Weaver, Krause, and Murray. Statistical analysis: Weaver. Administrative, technical, and material support: Gibson and Murray. Study supervision: Comfere and Gibson.
Financial Disclosure: None reported.